Exception routing calibration across systems: a key-by-key map of case, rule and review records

Short answer

Four joins rebuild one calibrated case: case to the rule that sent it somewhere, case to the response clock it was given, case to each reviewer's label, and the reviewers' ruling to the edit it caused. A shared case ID carries the first three; a session ID cited in a change ticket carries the fourth. Expect three systems, sometimes four. The case side is usually intact. The links into the rule log and the label sheet are the ones that fail.

Visual briefFull-size diagram
Three stacked system nodes on the left (case manager, routing tool and QA tool) send linking keys such as case ID, rule ID and version, and session ID across to a joined case ledger on the right, whose rows fill in one by one; one row is flagged where a spreadsheet is keyed by row number.
Illustrative. The provider, queues and identifiers are fictional and show how one case is joined across three systems.

This page maps those links for the exception routing calibration guide. A workflow automation advisor can use it in one meeting with the client's support, engineering and quality leads: name each system, the identifier it stores and the person who answers for it, then mark each link as intact, partial or missing.

Three systems, one calibration loop

Calibration runs in a loop. Cases get routed, a sample is pulled, reviewers label it blind, a lead settles disagreements, and an engineer edits a rule. Each turn of the loop writes to a different place, and each place uses its own identifiers.

  • Case manager. The ticket itself: text, channel, timestamps, queue, priority and a trail of field edits. Many platforms keep this trail in detail; Zendesk's ticket audits, for example, log every update with change events showing a field's previous and new values. Identifier: case ID.
  • Routing tool. The routing logic: ordered rules, their versions and, on good platforms, a per-case record of which rule acted. Identifier: rule ID plus version, and ideally the case ID in each log line.
  • Calibration store. A QA product or a shared spreadsheet holding sessions, blind labels, comments and the agreed answer. Identifier: session ID, plus the case ID on each label row if someone thought to add it. Some products mark the agreed answer explicitly; Zendesk QA, for one, lets a session flag a baseline review.
  • Change log. Often a ticketing project where engineers request and approve rule edits. Identifier: change ticket ID, which may or may not cite the session.

System, record, linking key and owner

SystemRecordLinking keyTypical owner
Case managerCase with text, channel and timestampsCase IDSupport operations manager
Case managerField change history: queue, group, priorityCase ID and audit timestampPlatform administrator
Case managerSLA policy and target applied to the caseCase ID and policy nameSupport operations manager
Routing toolRule set and its order on a given dateRule ID, version and effective dateWorkflow engineering lead
Routing toolExecution log: which rule acted on which caseCase ID and rule ID with versionPlatform administrator
QA tool or shared sheetCalibration session and sampleSession ID and case IDsQuality operations head
QA tool or shared sheetIndividual reviewer labels and commentsSession ID, case ID and reviewerQuality operations head
Change logRule change with commentChange ticket ID citing the session IDWorkflow engineering lead

Two rows carry most of the risk: the per-case execution log, because it may be kept only for a short window, and the label rows, because shared sheets rarely carry the case ID.

One fully joined case (Illustrative)

A fictional 450-person broadband provider runs its own customer service team. Two of its queues are Billing disputes (24-hour first response) and Retention (4-hour first response). A customer emails: "I was double billed again. Cancel my service." Nothing here describes a real company or a completed deal.

StageSystemRecordKey
ArrivalCase managerCase 884210 created by email, priority normalCase 884210
RoutingRouting toolRule R-14 v6 ("billed", "charge") assigns Billing disputesCase 884210, rule R-14 v6
ClockCase managerBilling policy, 24-hour target; first reply at 21 hoursCase 884210, policy Billing standard
OutcomeCase managerStatus solved; customer canceled two days laterCase 884210 audit events
SampleQA toolCalibration session CS-2025-07, case includedSession CS-2025-07, case 884210
LabelsQA toolReviewers: Retention, Retention, BillingSession, case and reviewer
RulingQA toolBaseline: stated intent to cancel overrides the billing topicSession CS-2025-07
ChangeChange logRetention rule R-09 moved above R-14; R-14 becomes v7Ticket citing CS-2025-07

Every row joins because the case ID appears in the router log and the QA session, and the change ticket cites the session. Take away the rule version in the router log and the routing step becomes a guess.

Where the joins break

JoinUsual breakQuestion to ask the owner
Case to rule firedThe router logs the queue assigned, not the rule that assigned itDoes the log name the rule and version per case, and for how long is it kept?
Case to SLA policy appliedThe policy order is edited in place with no historyIs there a record of the SLA policy order over time?
Case to reviewer labelLabels kept in a shared sheet keyed by row number or agent nameDoes every label row carry the case ID and date?
Ruling to rule changeRules edited with no reference to the calibration decisionDo change tickets cite the session or ruling?

A replacement project that migrates open cases and today's configuration but not the router log or the QA history keeps the outcome and loses the reasoning. The worked hard case shows what that reasoning looks like step by step.

What the map does and does not settle

A finished map shows which records exist, which years they cover, how they join and who owns them. It does not show whether they can be licensed. Case text written by customers falls under customer terms, so counsel shapes the answer; the who can license decision tree shows who decides. Platform background is on the Zendesk system page, and why operations data is valuable explains why the joined history matters more than any single export.

The introduced company needs 50+ full-time employees at peak (contractors excluded), documented operating history, rights to license and an authorized contact. Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 70% of the data is in English. What stays unknown until SourceX reviews it: rights in customer content, de-identification, SourceX's current data needs and any offer. Partners sign the Channel Partner Agreement before submitting, and the first valid introduction within the 90-day window is credited when the company submits a verified application.

Next step

Refer a company. No cap on rewards. Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

If a client's case IDs run through its router log and review sessions, register as a partner, confirm the company on who qualifies and read the program terms before you share the map with SourceX.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Company signs and is paidThe company signs, delivers its data and is paid by SourceX under its agreement.
  4. Step 4You get your rewardthe greater of $50,000 plus 5% and 15% of the actual supplier payout, with no cap. A reward becomes payable after SourceX has paid the referred company under its signed agreement and the payout conditions in the program terms are met, including finance approval and your tax and payment details.

Common questions

Do I need to know the vendor names of each system?

It helps, but categories are enough to start: case manager, workflow engine, QA tool or spreadsheet. The owner can add product names when they review your notes. Product names appear in a brief as plain text only.

What if the routing rules live inside the case manager?

Then the case-to-rule join may be internal, but ask whether the platform records which rule acted on each case or only the result. Ordered triggers and SLA policies in one product still need their history kept somewhere.

Can calibration labels in a spreadsheet be joined at all?

Only if each row carries the case ID and a date. Rows keyed by a row number or an agent name cannot be tied back reliably. Note which kind the client has; it changes how much of the history SourceX can use.

Sources

  1. developer.zendesk.com/api-reference/ticketing/tickets/ticket_audits
  2. support.zendesk.com/hc/en-us/articles/7043724530842-Setting-up-c…
  3. refer.sourcex.si/terms

Free resources

Know a company with valuable proprietary data?

Become a referral partner, get your link and introduce the company’s CEO (in an owner-run company, usually the owner). Referral partners can join from most countries. We can't work with partners in countries or regions under U.S. sanctions or on U.S. restricted-party lists. The companies you introduce should be based in the U.S., Canada and the EU, or in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia with at least 70% of their data in English.

Refer a company →

I own a business

Explore licensing your company's data directly to SourceX for AI training and evaluation. Start a short assessment; no uploads needed.

Start an assessment